The initial database search yielded 1414 results, and 883 studies remained after duplicates were removed. All 883 studies were then independently reviewed with exclusions based on abstract review. If any uncertainty remained regarding study eligibility after abstract review, the full text was reviewed. Studies were excluded for several reasons, the most common of which was that study patients were ≥ 10 years of age (n = 378). Other reasons for exclusion included: insufficient data (n = 231), textbook chapter (n = 92), not English (n = 96), not JNA (n = 53), and not human (n = 4). “Insufficient data” typically referred to studies that did not provide the age of included patients or that did not provide specific data for patients less than 10 years of age. At the conclusion of this process, 29 studies met our inclusion criteria [1, 2, 4,5,6,7,8, 21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46]. Figure 4 illustrates the details related to our study selection process.

Flowchart detailing search strategy and reasons for exclusion

A total of 34 JNA cases under age 10 at presentation were identified in the literature. Of these, 17 (50.0%) were 9 years old at presentation, 12 (35.3%) age 8, 4 (11.8%) were 7, and 2 cases aged 6 and 4. The average age was 8.15 years (SD = 1.08). Study designs included retrospective cohort studies (n = 12), case series (n = 10), case reports (n = 6), and 1 was a prospective cohort study. Patient follow-up time was reported in 19 of the included studies with an average of 32 months (SD = 41.42). There were five instances in which studies contributed multiple JNA patients under 10 years of age; [2, 23, 27, 29, 44]. The gender breakdown was overwhelmingly male in this cohort with 31 (91.2%) male cases and 3 (8.8%) female cases. Table 1 summarizes the study characteristics and patient demographics of the included studies.

Several presenting symptoms commonly associated with JNA growth were identified in this cohort. Symptoms were reported in 23 of 34 patients. The remaining 11 patients did not have data available regarding presenting symptoms. Patients most commonly presented with nasal obstruction (n = 15, 65.2%) and epistaxis (n = 14, 60.9%). Other associations included proptosis (n = 5, 21.7%), sleep disordered breathing (n = 4, 17.4%), and visual changes (n = 4, 17.4%). Ocular manifestations included decreased visual acuity, strabismus, and diplopia. Headache (n = 3, 13.0%) and facial swelling (n = 2, 8.7%) were also noted. Other nasal symptoms such as hyposmia (n = 1, 4.3%) and hyponasal speech (n = 1, 4.3%) were rarely reported in this cohort. An outline of presenting symptoms is provided in Table 2.

Overall stage at presentation was reported in 33 of 34 patients. As was previously discussed, the multitude of different staging systems creates a significant source of heterogeneity. The Radkowski and Fisch systems were most commonly utilized, in 6 and 4 cases respectively. Other staging scores used include Onerci, Andrews, Chandler, and Sessions staging. To minimize heterogeneity, primary staging for each patient was converted to a Radkowski stage using the original staging system presented or descriptions of tumor location and extension obtained from imaging or intraoperative accounts. Twelve patients (36.4%) were stage II at diagnosis, while 14 patients (42.4%) were stage III at diagnosis. Only 7 patients (21.2%) were stage I at diagnosis. The youngest patients, aged 6 and 4, were classified as stage III. Of patients that were 9 years of age, which was the most prevalent age in this cohort, 7 (43.8%) were stage II while 6 (37.5%) were stage III and 3 (18.8%) were stage I. An overview of stage at presentation vs. age is provided in Table 3.

Primary treatments reported in this cohort included surgery, radiation, or a combination of the two. This was outlined in 33 of 34 cases. Surgery was the most common primary treatment modality employed (n = 30, 90.9%). External surgical approaches such as lateral rhinotomy and transmaxillary approaches were cited most commonly in the surgical group at 66.7% (n = 22). In 3 external surgical cases, endoscopic assistance was utilized. A primary endoscopic approach was used in 24.2% (n = 8) of cases. Of note, 24.2% (n = 8) of cases managed with surgical excision required multiple operations, most commonly transitioning from an initial endoscopic approach to open approach. Postoperative radiation was given to 1 patient. Radiotherapy alone was used as a primary treatment modality in only 9.1% (n = 3) of cases. All three of these cases had intracranial and orbital involvement.

The most common complication reported in this cohort was excessive intra-operative blood loss (n = 9), which was defined as > 500mL. This occurred in 27.3% (n = 6) of patients that underwent primary open surgical excision and in 37.5% (n = 3) of patients that underwent primary endoscopic surgical excision. Blood transfusion was required for 13.6% (n = 3) of patients that had open surgery and 12.5% (n = 1) of patients that had endoscopic surgery. All patients that required blood transfusion lost ≥ 1L intra-operatively. Other complications reported in the non-endoscopic surgical group included post-operative trismus (9.1%, n = 2), cavernous sinus injury (9.1%, n = 2), lateral rectus paralysis (4.5%, n = 1), and hyperlacrimation (4.5%, n = 1). No additional complications were reported for patients that underwent endoscopic surgical excision. Cataracts were reported as a complication in 100% of patients (n = 3) that underwent primary radiotherapy. Growth retardation was reported as a complication following radiotherapy in 1 subject. An overview of reported complications and their relation to treatment modality is outlined in Table 4.

Table 5 illustrates the distribution of treatment modalities by Radkowski staging. Data regarding both treatment and staging were provided for 31 of 34 patients. Patients with stage I disease (n = 7) were managed with open surgery in 57.4% (n = 4) of cases and endoscopic surgery in 42.6% (n = 3) of cases. Of patients with stage II disease (n = 10), 70.0% (n = 7) were managed with open surgery, 30.0% (n = 3) with endoscopic surgery. Of patients with stage III disease (n = 14), 64.3% (n = 9) were managed with open surgery, 14.3% (n = 2) with endoscopic surgery, and 21.4% (n = 3) with radiotherapy. Follow up and assessment for tumor recurrence was reported in 28 of 34 patients. Average follow-up time was 32 months (SD = 41.4%). Recurrence was not reported for radiotherapy because this is not a curative treatment. Recurrence of tumor was identified in 28.6% (n = 9) of patients. Of those, 37.5% (n = 3) were following an endoscopic surgical approach and 25.0% (n = 5) were following an open surgical approach. The difference in recurrence rates between open surgery and endoscopic surgery was not statistically significant (p = 0.334, 2 = 0.933). In addition, recurrence risk was also analyzed in association with the primary Radkowski stage. 33.4% (n = 7) of stage II and III patients recurred, whereas only 14.3% (n = 1) of stage I patients recurred. However, this difference did not reach statistical significance (p = 0.595, 2 = 1.037).